ASTAM: Credibility is a model for within- and between-risk variation

Bühlmann credibility balances observed experience against a collective mean. EPV captures process variation within a risk; VHM captures variation in risk means. A large sample alone does not establish credibility without those components and model assumptions.

Worked example or practice scenario

For n=4 observations, EPV=100 and VHM=25, K=4 and Z=n/(n+K)=.5. If the risk mean is 120 and the collective mean is 100, the credibility estimate is 110.

Try this next

Double n and compute Z=8/12, giving 113.33. Explain why larger VHM would increase credibility with EPV fixed: the risks are more heterogeneous, so the risk’s own experience is more informative.

Reading sources

• Official syllabus and resources

ActNet editorial guide · October 1, 2026 · Original illustrative examples.